AI Paper Atlas

Research map / Learning theory and algorithms / Kernels and regression theory

On the Optimal Weighted $\ell_2$ Regularization in Overparameterized Linear Regression

Denny Wu, Ji Xu

NeurIPS 2020 · Learning theory and algorithms › Kernels and regression theory · approach: Kernel ridge regression analysis · 17 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Use a negative ridge parameter when feature variance and signal strength are positively aligned in the noiseless overparameterized regimeStandard ridge regression · Overparameterized linear regression, noiselessprediction riskImproved
Use a positive ridge parameter when feature variance and signal strength are negatively alignedStandard ridge regression · Overparameterized linear regression, noiselessprediction riskImproved
Increase the noise level in a setting where the optimal ridge parameter is negativeOverparameterized linear regression · Linear regression with anisotropic features and signaloptimal ridge parameterMixed
Use positive rather than negative ridge regularizationRidge regression · Underparameterized regime, γ<1prediction riskImproved
Increase the retained principal-component fractionPrincipal component regression · Overparameterized regime, θγ>1, with decreasing E[g|h]prediction riskImproved
Use weighting matrix Σ_w=Σ_x rather than another admissible weightingRidgeless generalized ridge regression · Overparameterized linear regressionvarianceImproved
Use weighting matrix Σ_w=Σ̄_β^{-1} rather than another admissible weightingRidgeless generalized ridge regression · Overparameterized linear regressionbiasImproved
Use weighting matrix Σ_w=Σ̄_β^{-1} with optimally tuned λOptimally tuned generalized ridge regression · Overparameterized linear regressionprediction riskImproved
Use a weighting matrix based only on Σ_x, proportional to (f_v(Σ_x))^{-1}Optimally tuned generalized ridge regression · Overparameterized linear regression when signal covariance is unknownprediction riskImproved
Use the minimum weighted-norm solution instead of PCRRidgeless generalized ridge regression with Σ_w=Σ̄_β^{-1} or (f_v(Σ_x))^{-1} · Noiseless overparameterized linear regressionprediction riskImproved
Increase the overparameterization ratio γOptimally tuned standard ridge regression with isotropic signal prior · Linear regression with general data covarianceprediction riskHurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy